Abstract: A computer-implemented method for minimizing the amount of time to access current data in a database which may be stored wholly in a DASD-oriented external storage subsystem or partly in DASD and partly in a stored high-speed electronic store while maintaining coherency of the data with respect to multiple user systems.

Abstract: Enhanced data availability occurs in a write-ahead logging, transaction-oriented database system by permitting new transactions to acquire access to data while restart recovery operations are proceeding. The invention permits new transactions to acquire access to data during restart recovery UNDO processing on the condition that the last update to the data occurred before a commit point measured by the beginning of the earliest-commencing transaction with uncommitted updates which was still executing when a system failure initiated restart recovery operations. During REDO processing, a transaction is permitted access to data which, in addition to meeting the commit point condition, is not in a data structure subject to the REDO processing.

Abstract: A fast technique for transferring units of data between transaction systems in a shared disk environment. The owning system, having updated the page, generates a version number for the page which is stored with a lock possessed by the owning system. When a requesting system seeks a record on the page, its request for a lock illicit an indication that a more recent version of the page is required in the local memory. The buffer management component of a DBMS, with assistance from the lock management, triggers a memory to memory transfer of the page from the owning DBMS to the requesting DBMS using a low overhead communication protocol. The transfer of page is without disk I/O or the log I/O for the updates made to the page.

Abstract: A high-speed cache is shared by a plurality of independently-operating data systems in a multi-system data sharing complex. Each data system has access both to the high-speed cache and the lower-speed, secondary storage for obtaining and storing data. Management logic and the high-speed cache assures that a block of data obtained form the cache for entry into the secondary storage will be consistent with the version of the block of data in the shared cache.

Abstract: In a multi-system data sharing complex, a database system can write updated pages to a shared electronic store for a fast write. Other database systems can obtain pages written to the shared store for further modification without the pages first being written to stable storage. However, pages are eventually written to the stable storage in a castout process. Recovery of a database from failure of the shared store is bounded by determination of a recovery boundary which, when applied to the union of database system transaction logs, establishes a point in front of which are found log records of modifications to pages which were in the shared store when it failed. These log records are applied to page versions obtained from stable storage to recover from failure of the shared store.

Abstract: An improved concurrency control system for application to a distributed concurrent transaction and query processing system using multi-version database records to overcome delays arising from lock conflicts. Read-only queries are afforded a consistent "stable state" of the database during the life of the query. Updating transactions requiring locks can proceed without waiting for the termination of long queries. At least two database versions are necessary, although availability of more versions permits long read-only queries to phase-out over time without forcing new queries to use aged "stable-state" data and without roll-back. Read-only queries can be terminated and converted to locking transactions to permit an update of the "stable state" database version before the queries would normally terminate. A novel record key structure having a plurality of substructures corresponding to the several database versions is used to access database records.

Abstract: A method of controlling entry of a block of data is used with a high-speed cache which is shared by a plurality of independently-operating computer systems in a multi-system data sharing complex. Each computer system has access both to the high-speed cache and to lower-speed, upper-level storage for obtaining and storing data. Management logic in the high-speed cache assures that the block of data entered into the cache will not be overwritten by an earlier version of the block of data obtained from the upper-level storage.

Abstract: A method of reducing the number of messages required for sync point (commit or backout) operations by leaving out nodes that have not participated in the corresponding transaction. A two-phase sync point protocol is used in a distributed transaction processing network to commit or backout transactions in the network. In response to the beginning of sync point operations on a transaction x, each node determines if each of its partner nodes stated on the sync point operation for transaction x-1 that the partner could be left out of sync point operations for transaction x-1. If a partner node did so state that it could be left out, the present node determines if the partner node was included by the present node during the present transaction x. If the partner node was not included during the present transaction, the present node excludes the partner node from the present sync point operations.

Abstract: Apparatus and method for reading data pages in a transaction processing system without locking the pages are disclosed. The system maintains a Global.sub.-- Committed.sub.-- LSN identifying the oldest uncommitted transaction accessing any of the data, and Object.sub.-- Committed.sub.-- LSNs identifying the oldest uncommitted transactions accessing particular files, tables and indexes. Each data page includes a Page.sub.-- LSN identifying the last transaction to have updated the page. To read a page, a transaction first latches the pages, and compares the page's Page.sub.-- LSN with the Global.sub.-- Committed.sub.-- LSN, or with the page's respective Object.sub.-- Committed.sub.-- LSN. If the Page.sub.-- LSN is older than the Committed.sub.-- LSN with which it was compared, then the transaction reads the page without locking it, since there can be no uncommitted transaction in process which might have updated the page's data. However if the Page.sub.-- LSN is younger than the Committed.sub.

Abstract: Results of a relational data base management system are joined in a process requiring, first, existence of an index on the join columns of the inner table, and, second, ordering on the join column of the first table. First, the index on the inner table's join column is scanned for rows of the inner table having join column values matching such values of rows in the outer table. This is done in a single pass through the outer table. Next, a temporary work table containing the identifiers of inner table rows having join column values matching those of the outer table is produced by concatenating the row identifiers to their matching outer table rows. Following this, the temporary work table is ordered by the identifiers. Last, the identifier list of inner table rows is used to retrieve the corresponding rows of the inner table.

Abstract: Change processing of a replica database is accomplished by separating redo records obtained from the transaction log of a primary database into respective queues. The redo records are separated such that all transaction records for a unit of transfer (page) of the primary database are placed on the same queue in log sequence. Each queue is linked exclusively to one of a plurality of parallel queue servers. Each queue server applies to the replica database the redo records in the queues which it exclusively serves. The replica database is thereby made consistent with the primary data by a lock-free updating mechanism which services the pages of the replica database in parallel.

Abstract: A method and apparatus for concurrent modifications of an index tree in a transaction processing system. The index tree includes at least one root node having a key record reference to one or more nodes in a next lower ordered level and at least one bottom node providing access to key records. Transactions including a structure modification operation are performed by traversing the index tree to the selected node and then setting an indication of the pendency of a structure modification operation. Concurrent key record inserts or deletes are permitted throughout the index tree where no indication of a pending structure modification operation is present and are delayed where a pending structure modification operation is indicated. Similarly, transactions which include a key record delete may require a structure modification operation in the event the transaction does not reach new point of consistency and must be undone.

Abstract: A method for fetching key record data in a group or record keys according to at least a portion of a key record through an index tree is provided. The index tree provides concurrent accesses of record keys by different transactions. The index tree includes a root node connected to at least one level of nodes, each node having a key record reference to one of more nodes in a next successive level and having bottom nodes that provide access to the key data.